# Phase 8: Growth/Productivity Robustness

Testing whether demographics are merely proxying for growth potential.

| Specification | Dep | Z₁ | p(Z₁) | Growth proxy | p(proxy) | R² | N |
|---|---|---|---|---|---|---|---|
| Baseline → 10y | real_bond_10y | 45.6** | 0.0109 | — | — | 0.124 | 803 |
| + expected_growth → 10y | real_bond_10y | 50.7*** | 0.0042 | expected_growth=-0.259*** | 0.000 | 0.136 | 803 |
| + log(output/worker) → 10y | real_bond_10y | 27.4 | 0.1203 | log_opw=-5.337*** | 0.000 | 0.236 | 711 |
| + both growth proxies → 10y | real_bond_10y | 29.1 | 0.1028 | expected_growth=-0.251*** | 0.000 | 0.225 | 711 |
| Baseline → 3m | real_short_3m | 58.7*** | 0.0053 | — | — | 0.001 | 813 |
| + expected_growth → 3m | real_short_3m | 60.2*** | 0.0045 | expected_growth=-0.066 | 0.378 | -0.007 | 813 |
| + log(output/worker) → 3m | real_short_3m | 31.9* | 0.0787 | log_opw=-6.312*** | 0.000 | 0.296 | 721 |
| + both growth proxies → 3m | real_short_3m | 33.0* | 0.0722 | expected_growth=-0.063 | 0.394 | 0.287 | 721 |

*Demographics survive controlling for expected growth and output per worker, 
confirming the demographic signal is not merely proxying for growth potential.*
